[NE Coffee Chat] Cynthia Produces Audiobooks in 1 Hour with AI That Normally Takes a Month

Interview with CEO Cynthia Oh Jin-hwan
AI Audiobook Auto-Production Service Targeting Global Market
75,000 Audiobooks to Be Produced Globally Next Year

About 4 million books are published annually in the United States, but only around 80,000, or 2%, are made into audiobooks. This is because it takes a voice actor three to four days just to read a 300-page book, and usually about a month to produce the audiobook. Production costs also run into hundreds of thousands of won. This is why only books with proven sales are turned into audiobooks. Cynthia is a startup aiming to revolutionize this inefficient audiobook production method using large language model (LLM)-based artificial intelligence (AI) technology. The technology they have honed since their founding in 2020 is expected to shine in 2025, when the global AI audiobook market fully opens.

On the 26th, CEO Oh Jin-hwan of Cynthia said, "The audiobook market has been thirsty due to a lack of supply," adding, "We are targeting the 98% of books that are not made into audiobooks based on the U.S. market." Cynthia's AI technology completes audiobook production, which traditionally takes 3 to 4 weeks with voice actors, within an hour. It is not just about converting text to speech (TTS). The AI first reads and analyzes the entire book before narrating it. CEO Oh explained, "With technology specialized for audiobooks, the AI breathes, adjusts intensity, and expresses emotions to produce natural sounds. For novels with characters, it creates fitting voices."

The AI synthesizes the voice most suitable for narration after reading the book. Considering that each individual has a voice that enhances immersion, multiple voice versions can be created for a single book. For names or creations made by the author, the AI applies the author's vocal characteristics to the narration voice, producing sound as intended by the author.

Cynthia CEO Oh Jin-hwan is explaining the AI audiobook automatic production service at the startup competition pitching event. Provided by Cynthia

CEO Oh explains that applying this solution allows anyone to create high-quality audiobooks affordably. Cynthia’s tentative pricing model is $9.9 per hour. For a 10-hour audiobook, it would cost only $99. He said, "We want to provide opportunities to authors who couldn’t produce audiobooks due to high costs and offer readers a broader, more accessible way to consume content."

Having built expertise in global business development by leading overseas operations at ST Unitas and co-founding the global Korean language education platform eBridge World, CEO Oh primarily targets the U.S. and European markets. Although Cynthia offers 19 languages including Korean, the English-speaking market is prioritized due to higher LLM application completeness. He said, "In B2B, there are many potential clients such as major publishers, audiobook platforms, music platforms, and production companies," adding, "As AI audiobooks increase on music platforms, subscription models will become established."

However, competitors are formidable. Global companies like Amazon, Apple, and Google have entered this market. From next year, AI-generated audiobooks from these companies are expected to flood the market. These global companies distribute AI-converted audiobook content exclusively through their own platforms. Cynthia sees potential by targeting this point. CEO Oh said, "Global companies have a closed model that monopolizes quality books and authors and distributes profits, so if authors want autonomy and to upload audiobooks on various platforms, it doesn’t fit. Using Cynthia’s service, they can create audiobooks cheaply and freely, and authors retain ownership." He believes that even if global companies take 50-60% of the market, there is ample room for growth in the remainder. He said, "Our goal is to secure the largest market share outside the 50% held by major players, and we expect 75,000 audiobooks to be created globally using our solution next year."